Subject: Re: KA6x0 support
To: None <port-vax@NetBSD.ORG>
From: Michael Sokolov <msokolov@blackwidow.SOML.CWRU.Edu>
List: port-vax
Date: 06/18/1998 22:42:54 (Ty Sarna) wrote:
> Excuse me? Have you ever looked *outside* it?
   No. Don't forget that the ONLY reason I pay attention to NetBSD and
subscribe to this list is because I want to exchange ideas and sometimes
even code with the NetBSD gang in the area of VAX HARDWARE SUPPORT.
> Excuse me? i386er? I run NetBSD on some pc boxes now, but I've been
> using NetBSD since late 0.8 or so, and it wasn't until after 1.1 that I
> even booted it up on a i386.
> Don't call me a "new hardware' bigot, either.  I'm not running any
> VAXen, but I do have a production machine (and still a fairly important
> one at that) with a manufacture date that's almost a decade old.
   In order to understand the problems with the current implementation of
uVAX support in NetBSD you really need to know VAX hardware at the level of
chips and registers.
> You were claiming
> here, and in previous messages months back, that NetBSDs config
> mechanism was broken. Not NetBSD/vax's use thereof, but NetBSD MI
> config. You said quite pointedly that New Config (which you don't know
> anything about, obviously, other than that it isn't as familiar to you
> as Old Config) was a crock.
   This, as well as all my other OS choices, is a personal preference that
I choose not to discuss on this list. Before my 3-month leave from this
list I was making the mistake of discussing it here. I now realize that
this was inappropriate and refrain from discussing my OS preferences on
this list.
> Excuse me? I can probably dig up a copy of a message from you to
> lynx-dev where you described NetBSD as "amazingly bad", or words to that
> effect. I don't recall wether that was qualified as NetBSD/vax or not,
> but it was unfair nonetheless.
   Yes, I remember that post. It referred to NetBSD/vax specifically. At
that time I was working on my KA42/41 SCSI fix, but not being able to
netboot I was unable to get the thing installed in the first place! (I
obviously needed a NetBSD/vax installation in order to recompile the
kernel, but since I was unable to netboot and the stock kernel can't talk
to SCSI devices on KA42/41 AT ALL, I had no way of installing it.) I did
refer to NetBSD/vax as being "unbelievably crappy" in that post, but how
else was I supposed to describe an OS that was _impossible to install_? In
any case that problem is now fixed, by no one but me. (The reason I was
posting that to people on the net was because I was looking for a volunteer
with a NetBSD or FreeBSD system to mount the miniroot and replace the
kernel with the PIO one. I had no luck with that, but I was saved by Bruce
Richards, the manager of another department and a friend of mine, who has
graciously allowed me to borrow his VS3100 M76 for a week, making it
possible for me to bootstrap NetBSD by playing the move-the-disk game.)
> Excuse me? You keep your preferences to yourself? Since when???
   Since my return to this list from a 3-month leave.
> Excuse me? You don't even know what the truth is.
   I do know it. The truth is that the uVAX code in NetBSD has both
strengths (like its very existence) and weaknesses (poor organization and
extendibility). I'm not using the latter, however, to depict NetBSD as an
inferior OS, since no other version or flavor of VAX UNIX can currently
boast having better uVAX support. (As I've said before, the uVAX code in
Ultrix is also flawed. The Ultrix designers have made several fatally wrong
decisions early on, like supporting KD32 aka MV I and viewing on-board
devices at being attached to uba0, that preclude Ultrix from ever running
on KA45 and sons, even if someone with the full sources tried to do
something about it. Of course one could totally restructure the code, but
then it would no longer be Ultrix.)
> Excuse me? If it's based on 4.3BSD, how is it 100% your own?
   I have meant that the _uVAX support code_ will be 100% my own.
> You're giving us orders of magnitude more attitude, and expect us to put
> up with that for what? *Three* lines of code?
   Sigh. You still don't understand. Attitude is precisely what I have to
offer! My 3-line patch has been intended specifically as a spectacular
demonstration of the fact that many of NetBSD/vax's major problems owe
their existence not to the lack of code "meat", but to misguidance. What my
patch has corrected was exactly that: an ideological misguidance (the
placement of a particular system board in the wrong class). I also believe
that I have already explained to death by now that my attitude and my
unconventional methods will be the primary means by which I will approach
my very ambitious goal of supporting nearly every VAX from 780 to 10000.
   Michael Sokolov
   Phone: 216-368-6888 (Office) 440-449-0299 (Home) 216-217-2579 (Cellular)
   ARPA Internet SMTP mail: